Brain tumors are the most common solid tumor of childhood. Some can be cured, but for the 30-35% of pediatric patients who recur following front-line therapy ? and for every child with diffuse intrinsic pontine glioma (DIPG) ? there is essentially no chance of cure with standard treatment. Thus, more effective therapies are urgently needed. The scientific premise underlying the current proposal is that conventional radiation and chemotherapy release large amounts of antigen from dying tumor cells, but that this normally cannot trigger a useful immune response because immunogenic antigen-presentation is suppressed by tumor-induced mechanisms such as the indoleamine 2,3-dioxygenase (IDO) pathway, a natural mechanism that profoundly inhibits immune response to apoptotic cells. The proposal hypothesizes that adding immunologic therapy using indoximod, an oral inhibitor of the IDO pathway, in combination with radiation and chemotherapy, will allow prolonged survival in these otherwise refractory patients. The applicants have recently completed a first-in- pediatrics Phase 1 study of the proposed indoximod-based chemo-radio-immunotherapy approach in 53 children with recurrent brain tumors. This proof-of-concept study shows a median Overall Survival of 17.2 months, which is markedly superior to historical comparator studies, with low toxicity. Aim 1 will conduct a Phase 2, stratified-design trial of indoximod immunotherapy in combination with temozolomide (TMZ) chemotherapy, with or without the addition of novel Low-Dose Partial-field (LDPF) radiation, in 91 pediatric patients with recurrent ependymoma, medulloblastoma and GBM. Outcome measures will be 8-month Progression-Free Survival (PFS), and 18-month Overall Survival (OS) and Time to Regimen Failure (TTRF), as compared to PFS and OS from historical controls. Aim 2 will conduct a Phase 2 single-arm trial of 30 children with newly-diagnosed diffuse intrinsic pontine glioma (DIPG), to test the hypothesis that front-line indoximod added to standard conformal radiotherapy, followed by maintenance chemo-immunotherapy with indoximod + TMZ, will result in improved Overall Survival (OS), compared to well-documented historical controls. Aim 3 will use mechanistic predictions from novel preclinical models to identify innovative, hypothesis-driven immune biomarkers, combined with cell-intrinsic genomic and epigenomic features of the patients' original tumors, to ask whether these allow prognostic risk stratification of patient outcomes in Aim 1 and Aim 2. A successful outcome from the proposed clinical trials has the potential to fundamentally change the approach to treating children with recurrent or refractory brain tumors. It would also have major implications for incorporating immunologic therapy directly into the standard treatment for children with high-risk brain tumors, at the time of front-line treatment, when there is the real possibility of long-term cure.